Adam and Vitalik. The future doesnt need you.
SingularityNET CEO on the ASI Alliance and the future of decentralised AGI
Illustration: Andrés Tapia; Source: SingularityNET.
HomeSingularityNET CEO on the ASI Alliance and the future of decentralised AGI
By
DL Research
27 January 2026
Dr Ben Goertzel is a leading figure in artificial general intelligence, holding a PhD in Mathematics, having authored over 25 books, and publishing more than 150 research papers.
He heads SingularityNET, TrueAGI Inc., the OpenCog Foundation, and the AGI Society, and previously served as Chief Scientist at Hanson Robotics, where he contributed to the development of robots such as Sophia, Desdemona, and Grace. His work on OpenCog Hyperon and decentralised AI infrastructure has made him one of the most influential voices advancing open, ethical AGI.
We recently spoke with Dr Ben Goertzel, CEO and Chief Scientist at SingularityNET and CEO of the Artificial Superintelligence (ASI) Alliance, about the technical architecture behind ASI:Chain and how it differs from existing blockchain infrastructure.
Read more about the convergence of blockchain and artificial general intelligence in the interview below.
ASI:Chain starts from the idea that a blockchain should natively support distributed intelligence. What core technical requirements of autonomous agents made it impossible to build this vision on existing L1s and L2s?
Agents need a chain that can natively handle massive parallelism, AI compute and memory primitives, micropayments and dispute settlement, privacy-preserving trust and governance, and heterogeneous execution and consensus.
The blocker for existing blockchains is not one feature but rather the fact that agents need the chain to behave like an AI operating substrate that handles compute, memory, coordination, payments, and governance, not just serve as a settlement ledger.
In other words, AI agents require blockchains that function as complete operating systems for artificial intelligence rather than simple transaction ledgers.
Hyperon Alpha 1 introduces scalable Atomspace and a production-ready MeTTa engine. How do these upgrades change what developers can practically build today compared to earlier AGI research frameworks?
It turns “AGI architecture” into something you can actually deploy. MORK makes Atomspace fast and concurrent, and DAS scales that same “brain space” across clusters. Meanwhile, MeTTa and MeTTa-IL provide a clear, compilable, verifiable path to multiple backends so developers can build real systems that share intermediate cognitive state, not just pass API calls around.
This stack transforms AGI from a theoretical framework into a practical development platform where systems can share actual cognitive structures instead of just exchanging messages.
MeTTa will serve as a core smart-contract language for ASI:Chain. What does “reasoning-capable smart contracts” unlock that Solidity-style contracts could never support?
MeTTa contracts enable blockchains to execute actual reasoning and decision-making processes on-chain instead of just storing and updating data.
MeTTa contracts can operate on knowledge graphs and reasoning processes directly (pattern-match, rewrite, explore alternatives, then “commit” deterministically) rather than only updating storage variables. With committed-choice evaluation and transactional Space operations, you can encode negotiation, planning, evidence-based decisions, and safe multi-step cognition as on-chain logic.
ASI:Chain uses a blockDAG and configurable multi-consensus shards. How does this architecture handle billions of concurrent agent interactions without compromising determinism or security?
The system scales through two main mechanisms: parallel execution via a blockDAG structure and horizontal scaling via shards that can each choose their own consensus algorithm based on their workload.
MeTTa’s design guarantees that operations produce consistent, predictable results through features like committed choice and atomic Space operations. Security is maintained through shard-level controls, such as capabilities, access restrictions, and isolation, combined with fault-tolerant consensus mechanisms, such as DPoS and CBC Casper, in the DevNet implementation.
What kinds of AI-driven applications do you expect to migrate to specialised shards that use different consensus mechanisms based on their workload? Where do you see the earliest real-world demand for open, decentralised agent coordination on ASI:Chain?
The first movers are workloads that are compute or coordination-heavy, such as decentralised compute and inference markets, shared-memory multi-agent systems, real-time simulations and world models, and privacy-first “compute-to-data” apps.
Early traction is expected around working compute and memory shards with stablecoin payments in a multi-agent environment and practical dispute settlement. In other words, “open agent coordination that actually ships.”
Hyperon provides a shared “brain space,” while ASI:Chain provides the decentralised spine linking those systems together. How do you see this stack reshaping the balance of power between closed AI labs and open communities?
This stack reshapes the industry by decoupling contribution from centralised control. Hyperon provides a modular brain space, with the “MettaCycle” framework ensuring security and provenance. ASI:Chain then adds the economic layer, enabling open contributors to coordinate and monetise their work without depending on the proprietary rails of hyperscalers.
If the next wave of innovation toward AGI happens on this decentralised network rather than inside proprietary labs, it will fundamentally shift the global power balance.
This structure undermines political leaders’ ability to frame AI progress as a zero-sum geopolitical “AGI race.” Instead, it positions AGI as a global commons like the internet or Linux. This is the necessary path as we prepare for the transition to superintelligence and the profound changes it will bring.
One of the Alliance’s stated goals is ethical and democratic AGI. How does ASI:Chain’s design support that philosophy in practice?
Two big levers: verifiability and governance. MeTTa-IL is designed to make behaviour unambiguous so properties can be verified before deployment, while the chain and economy design include governance safeguards such as time locks and optimistic protections, as well as on-chain democratic mechanisms like DPoS.
In essence, safety comes from being able to verify what code will do before it runs and having democratic governance structures that prevent malicious changes.
Looking ahead to Testnet and Mainnet, what breakthroughs do you expect to emerge when Hyperon-based cognitive systems begin operating natively on ASI:Chain rather than in isolated environments?
We are moving from isolated demos to real, composable agent economies. In this environment, agents pay for compute and memory, share intermediate cognition, and coordinate at scale using native sharding.
The roadmap outlines a concrete near-term breakthrough: compute and memory shards operating in a multi-agent environment powered by stablecoin payments. Hyperon adds the necessary layer of provenance and secure multi-party cognition on top.
This translates into tangible applications across every vertical market. We expect to see assistants that retain long-term memories to help users pursue long-term goals, humanoid robots capable of navigating complex environments such as hospitals, and financial engines that adapt to different market regime
No comments:
Post a Comment